Okay , to recap ...

--I expressed some grave concerns that the admins of this site would be letting themselves in for endless headaches trying to find a single font they could use for displaying Traditional Hanzi in a manner consistent with the Heisig/Richardson book. (nb: the differences are not vast, but the handful that exist really throw a monkey-wrench into the Heisig method).

--I discovered (to my delight) that it was indeed possible to mix and match; to specify an overall font that closely matches H/R, and then slip in a few changes when appropriate.

-- I also discovered (much to my incredible delight) that I could also do this in ANKI, via CSS.

To cut to the chase, I am now methodically going through the first 1,500 trads to try to identify the optimum freely available font to use. I'm testing both MING and KAI styles (whichever style the admins ultimately goes with is entirely up to them. I'm indifferent on the matter. My own personal database will have both anyway. I would, however, like to provide them with some information that would make the choice easier. I would also suggest that they don't commit until they see the results).

That leaves me to ask: who exactly is making the design choices and doing the final coding?

ファブリス I would assume.

Still plugging away.

Working on a unified Ming list first. Will tackle Kai shortly.

Found one freely available Ming font that covers the vast majority of the hanzi quite nicely. Unfortunately, I still need to use 3 or 4 other fonts to fill in the gaps. And there are 1 or 2 trouble spots -- characters that don't exist in any font I have on my system.

That raises a larger issue that's troubling me. If I have close to 40 Chinese fonts from a wide variety of sources (free, commercial, from popular repositories and obscure Chinese sources) coupled with another 20 or so Japanese fonts -- and they say that a give character looks looks like "X", how can Heisig/Richardson present it as "Y" in their book? I can understand a few common variants; but some of these are really obscure in the real world. Beyond the (not insignificant) annoyance of trying to track down a font for them -- I gotta wonder at the value of committing them to memory in the first place.

Not like it will matter much since it's a website. What fonts Fabrice use is quite irrelevant, the fonts installed by the user using the site is where it matters.

Tobberoth Wrote:Not like it will matter much since it's a website. What fonts Fabrice use is quite irrelevant, the fonts installed by the user using the site is where it matters.

Not entirely so.

Fonts can be embedded via CSS.

Granted, I don't pretend to know the logistics nor do I know how they intend to approach it here, but I believe that was an idea raised. I don't know if there is a method to embe subsets or not. I sure hope so, or they're looking at pushing a minimum of 3 (and possibly 4) fonts on users.

Maybe someone with font software can concoct a single "Heisig Font" from spare parts.

I dunno ...

Heisig is giving me migraines.

Might I recommend emailing the publisher to see what font they used for the book?

mafried Wrote:Might I recommend emailing the publisher to see what font they used for the book?

Ehhh, they're all over the map. That's the problem. They chose characters that complimented their other books, sometimes (I fear) at the detriment to those interested in Chinese. And I understand their motivation and can't really blame them.

Now, I don't want to overstate this. Out of the whopping 1,500 from RTH, I only found (4) characters that I couldn't find an appropriate match for. Those are pretty good results, so I can't complain too much. And the differences are trifling to a degree, except that they invariably involve an established primitive that they are building on.

#0469 -- 說
#0689 -- 別
#1038 -- 麽
#1284 -- 絕

All the rest I can match pretty exactly -- but I need to pick and choose between (5) fonts. I'm trying to narrow that down, at the moment. If I recall correctly, the base font I chose handled all but apx. 145 characters out of the 1,500. And they were mostly dead-on matches.

Oh, and I'm only through the "Ming" version -- the print-style font. I've yet to compare the "Kai" calligraphic style. I have a pretty strong base font that seems to handle a great many. We'll see.

Anyway, to answer your question -- they use a proprietary typeface, according to Dr. Richardson, so we're SOL on that score. He did mention that there were a few Apple fonts (eg: Li Sung Light) that included the "correct" variant, but that doesn't do me much good.

Note that Heisig uses 麼 and not 麽. 麼 is much more common in traditional Chinese, and 麽 is not even present in the Big5 encoding (or maybe that's why it's less popular?). Actually, I think 麽 might be simplified Chinese.

Here are some confusing points about the difference between mainland and Taiwan usage:

Also worth noting is that 麼 is written as 么 in Simplified Chinese [me in shénme]. However, in Traditional Chinese this character is pronounced yāo (used e.g. for 1 when reading numbers), which in simplified corresponds to 幺 (maybe this character doesn't exist in traditional?).

So:
T<=>S
麼<=>么
那麼<=>那么
好麼<=>好么 (ma) also 好嗎<=>好吗

么<=>幺
么麼<=>幺麽

[ファブリス]

FuDaWei Wrote:Fonts can be embedded via CSS.

It works in modern browsers, but is still far from a standard feature, it's just about maturing now. Also fonts for chinese will be quite the download. Regular text will have to use a common chinese font that most people have installed on their OS or can find easily, and the Study area characters may use images (which themselves could be generated from a set of chinese fonts to get variances in characters).

ファブリス Wrote:

FuDaWei Wrote:Fonts can be embedded via CSS.

It works in modern browsers

Let me guess... Does it work in IE8?

ファブリス Wrote:

FuDaWei Wrote:Fonts can be embedded via CSS.

It works in modern browsers, but is still far from a standard feature, it's just about maturing now.

Just as a specialized site like this has the right to assume users have Chinese|Japanese capability installed on their system, I think you have the right to assume they'd be using a reasonably current version of a mainstream browser.

ファブリス Wrote:Also fonts for chinese will be quite the download.

What about the notion of cannibalizing the 1,500 (soon to be 3,000) Heisig characters we need and slapping them into a "custom" font, like the stroke-order font? Would that ameliorate the problem somewhat?

ファブリス Wrote:... the Study area characters may use images (which themselves could be generated from a set of chinese fonts to get variances in characters).

I've strongly suspected from the start that you'd be forced to resort to images. Hope some genius comes up with something.

In the meanwhile, I'll keep searching for the best font-combo. If nothing else, I can use the info in Anki.

Why not study them the way they are drawn in actual Chinese fonts? I.e, the way they will actually been seen and used in real life.

Am I missing something here?

[ファブリス]

Some further thoughts on this as I am browsing through the RevTH discussions.

I could generate images for the RTH/RSH characters. But I would prefer a long term solution. In the long term the site has also to support all standard Chinese characters, not just those that appear in the Heisig books. All characters in the database are uniquely identified by the UCS codes. EDIT: My point: too many images to generate or to cache efficiently by the browser.

So for the first version of RevTH I will use fonts, most likely.

What we could work out together is what would be the best font list to use, ie; a "recommended font list", that can be documented on the website, with links for users to download. Then I can reference these fonts by CSS, and if the user has them installed they will be properly selected.